Run parallel

Guest -

9:11:04 AM PST - Mon, Nov 8th 2010

I am running the pre-compiled NWCHem 6.0 Linux binary on a XEON machine with 4 processors and Suse Linux without problems in sequential mode. However, I have not been able to run it in parallel mode.
I created the nwchem.p file according to the intructions in the Documentation link, all the path stuff is apparently O.K. but any trial fails with error messages like that:

 sock=0, pid=1, nsent=-1, len=56

 1: snd_remote a: writing message to socket 39999999

 system error message: Bad file descriptor

  (and a few similar lines....)

The last entry in the output file has a message:

 3: ARMCI aborting 0 (0)

 Creating: localhost, host, file=  etc  with apparently correct entries

I suspect that I am lacking something to run parallel
Can somebody help ?

Thanks
Luis

Guest -

8:25:01 AM PST - Wed, Nov 10th 2010
Hi Luis, Can you send the command you are using to run in parallel ? Best, -Niri

Guest -

2:19:10 AM PST - Thu, Nov 11th 2010

Thanks, Niri

The command, run from directory /home/lfp/nwchem, is

bin/parallel bin/nwchem fileinp.nw > fileout.out &

I created the file nwchem.p in directory /home/lfp/nwchem/bin with the entry:

 lfp atenea 4 /home/lfp/nwchem/bin/nwchem /home/lfp/nwchem

NWChem starts, the output file is created (input info properly read, basis set stuff, etc correct) but the program aborts after a couple seconds and the last lines in the output file are:

3: ARMCI aborting 0 (0)

Creating: localhost=atenea, host=atenea, usr=lfp,

             file=/home/lfp/nwchem/bin/nwchem, port=9427

Thanks again
Best regards

Luis

Forum >> NWChem's corner >> Running NWChem